Dyno testing for Ebikes

Interesting conversation with a performance motorcycle shop here in T.O.

Their dyno (and I'm assuming others as well) can test down to 2 hp!!

It has an assist system to get the rollers going to help on the bottom end, then the test happens.

Cost for a dyno run is $65.

They claim a precision of 0.1 hp, but I wonder about that. I don't know if the individual I spoke to actually knew the error tolerance on the machine, but he was so fabulously helpful compared to the other fools I spoke to today, I was ecstatic just to talk to him.

Katou

Wow, the silence is deafening. Have I just discovered the modern day equivalent to building birdhouses for the Dodo?

Special jackets for your pet Great Auk?

Am I way off in thinking that this service would be useful to those tuners out there?

Katou

ps. me for one, would like to see the power difference between using one battery chemistry vs another.

Now I just need several battery packs of differing chemistry to try out. Anyone local want to volunteer?

Cool, that would be nice to know exactly how much power that goes into drivetrain etc.
To bad I don't live in the US. I guess we have the same dynos at home, but I bet its very pricey

We know that having high-C rated packs gets the juice to the motor when it needs it, and that high-C packs don't sag as much under load.

But is there a point of diminishing returns? Is there a point where having more C-rate is unnecessary?

Does powering a motor with a high-C pack equal the performance of just having a (low C) really big pack?

In theory, a 20C pack with 10 ah can put out 200a for some small period of time.

If we have a 2c pack, but 200 ah, it can also put out 200a

Would there be any difference in practice I wonder...

Katou

Datalogging would tell you that with more datapoints than a dyno.
And while a dyno may be capable of measuring down to 1500 watts with a resolution of 75 watts, it doesn't mean its going to be accurate at that extream edge of it's capability. It was designed for things atleast 10,000% more powerfull , and is ment to be accurate in that range usualy at 5%.

The shop near me's dyno will be off 2 or 3 horsepower on the same car in consecutive runs. The shop my friend used to go to with his honda said they consistantly got higher HP runs at the end of the day because the sensors read different at different temps.

There's simply no way something ment to be accurate for a car is going to be accurate for a bicycle.

You wouldn't try to accuratly measure baking ingrediants on a bathroom scale, would you?

If I was making a really, really big cake?

I know that accuracy is the problem. The problem specifically is knowing the accuracy. What is the tolerance? I don't know.

Could you explain using datalogging for this purpose? I don't know how to do that.

My next purchase is to be a datalogging RC controller, so I am quite interested in learning.

Katou

Oh, the dyno I was thinking of using was specifically designed for motorcycles, so that at least reduces the magnitude of the problem somewhat.

Katou

Horse power are Watts. 750 watts is 1 horsepower, so you can directly convert it. Horsepower is also torque over time. so if you have the voltage, amperage, speed, and time logged, you can get the torque and horsepower. you just need a constant form of measure. like a dragrace start for an 1/8th mile. The data logger is actualy just doing what the computer is doing to a dyno's drum, but with more input.

Ri.
Internal resistance.
If you know it for a cell, you can easily answer all those questions.

katou said:

We know that having high-C rated packs gets the juice to the motor when it needs it, and that high-C packs don't sag as much under load.

But is there a point of diminishing returns? Is there a point where having more C-rate is unnecessary?

Does powering a motor with a high-C pack equal the performance of just having a (low C) really big pack?

In theory, a 20C pack with 10 ah can put out 200a for some small period of time.

If we have a 2c pack, but 200 ah, it can also put out 200a

Would there be any difference in practice I wonder...

Katou

Click to expand...

I'm confused though, isn't the whole point here to do testing to see if it actually comes out that way at the rear wheel?

I'd also like to know how minor misalignments in the drivetrain parts (inevitable in home builds) result in losses at the rear wheel.

Seems to me that the only way to figure out how much of calculated hp is getting to the rear wheel is to do a dyno test.

Anyway, it seems like testing on a dyno is not as valuable as I thought. Oh well.

Katou


ps. LFP, if you have a thread or a post that explains RI better, please post it. I've read through Dr. Bass' stuff, and I still don't quite get it.

There are also some bicycle dynos out there that may be a better fit to ebike output power than a motorcycle dyno.

Really? Can you suggest where to look for that? I haven't had much luck in that direction.

Katou

katou said:

Really? Can you suggest where to look for that? I haven't had much luck in that direction.

Click to expand...

I'll see what I can find. I know a (now defunct) bike shop had one that was good to about 10 hp. I'm not sure if it was designed for bikes or some other application.

A motorcycle dyno should be a lot better.. the drum MOI should be 10x closer to what you need than a car chassis dyno. Motorcycles are like 400 pounds vs. 4000 pounds for a large car. Some small ICE motorcycles and scooter make less power than the higher power ebikes on ES. An ebike with a rider on board should be what 150 to 300 pounds vs. a rider on a motorcycle being 600 or so. Anyway a static mph/rpm test should get a pretty accurate reading I would think. One sure way to find out..

I have access to- and factory training on- a DJ 250i load control dyno, although experience tells me that the brake app window will ratchet under the tiny power of a bike (the 250i uses an eddy current bus brake). It'll be useful for finding saturation point, the place where diminishing returns are overwhelming, and setup housekeeping, like speed limiting and stuff, but honestly the dyno- unless you're actually winding motors or building drives- may not really tell us a whole lot we don't know- except what an x5 smells like molten!

I've put weak (ice) bikes on the dyno and it runs just fine and outputs data, so the biggest problem might actually be just setting the thing up for the proper wheelbase. The biggest boon might come from electrical analysis of the bike under load- especially controller torturing and finding out the rider's contribution at various speeds. Also, if you could set the brake up well enough, you could use it to simulate hills.

If anybody wants to do an ebike dyno day here in gorgeous Sonoma County, that would be fun. Also, if anybody wants to do dev work with the dyno, I can hook that up too for a charm.